Overall takeaways

What we saw and heard on the tour confirmed—emphatically—that students value their relationships with teachers and want to learn from them, even when they recognize that AI-powered tools can also play a role in facilitating their learning. The teacher-student relationship will undoubtedly continue to be central to students’ experience of school. 

Beyond just a relational role, teachers will continue leading the work of teaching children. Nothing we saw on the tour so far suggests that AI-powered instructional tools will be capable of taking the lead pedagogically (i.e., take the lead on all teaching and learning). Teachers and instructional leaders will remain the pedagogical authorities, holding responsibility for and leading the work of making sure that learning happens for every student.

There are AI-powered products and uses that offer students more practice, more consistent access to high expectations, and more feedback than we often see in teacher-delivered, paper-pencil instruction. We saw AI-powered products offer learning experiences that were more dynamic and engaging than those offered by edtech products of the past. We saw examples of well-designed products that attend to the hardest-to-deliver and most important drivers of learning (e.g., precise content questions, well-paced practice opportunities, opportunities for elaboration, high expectations held with just-right support). We also saw that products can improve at scale more quickly than we typically see teaching practice change. All of this suggests that AI-powered tools have an important role to play in the future of teaching and learning.

We saw AI-powered tools and uses that were no better than or underperformed typical teacher-delivered, paper-pencil instruction and learning experiences supported by edtech products of the past. In these learning experiences, we saw students go through the motions of learning without internalizing much—experiences that were isolating and boring to students. (Boring and isolating learning experiences can also be found in paper-pencil instruction, but the version we saw with AI products was even more dispiriting.) We saw kids tune out when they were asked to slog through lessons and practice problems that held little interest. AI-powered instructional products will only add value when they are carefully designed and implemented well.

We heard widespread concern from students and teachers about how students (and teachers) use general-purpose chatbots to get out of work that requires effortful thinking. Learning anything requires effort, and products that intentionally or unintentionally reduce students' effortful thinking will significantly compromise learning. It will take deliberate action (e.g., policies that limit access, conversations about academic integrity) to mitigate these risks.

Though teachers frequently indicated that they valued flexible products, we found that products with a clearer (and by definition less flexible) point of view about how they should be used supported stronger learning experiences that are likely to yield more consistent, positive student learning outcomes.